Philips alien-head LED lamps locked up during power brownout

PhotonWrangler

Flashaholic
Joined
Oct 19, 2003
Messages
14,432
Location
In a handbasket
I have two of the Philips remote phosphor "alien head" lamps in an overhead fixture. Yesterday we had a brief brownout and most of the lamps dimmed and then came back on, but the two bulbs in this one ceiling fixture went out and wouldn't come back on until I cycled the light switch.

There are no electronics in the fixture itself and the light switch is a regular, passive SPST on/off switch. Apparently something in the lamp driver stalled out. I doubt that this is the action of some sort of protection circuit but I'm not sure. The bulbs have been running fine since the brownout. I'm wondering what happened.

:thinking:
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
In my work life, I do some validation testing of electronics. Some of this includes slowly ramping up the supply voltage. There have been cases where semiconductors latch up due to junctions not getting reverse biased at the right time. Think of the substrate in an IC that is doped material, but must be reverse biased in order to avoid conducting. Well, once it starts conducting, it's not going to stop until you remove power.

I've also seen this sort of thing during ESD testing, for what it's worth. It's amazing what sort of weird stuff happens when you poke and taunt the electronics in just the right way. :)
 

PhotonWrangler

Flashaholic
Joined
Oct 19, 2003
Messages
14,432
Location
In a handbasket
I've also seen this sort of thing during ESD testing, for what it's worth. It's amazing what sort of weird stuff happens when you poke and taunt the electronics in just the right way. :)

A long time ago I was troubleshooting an AM radio that had intermittent sound. It turned out that a flaky transistor in the audio path was randomly acting as a switch. If I touched the base once with a screwdriver, it turned off. Another touch and it turned back on. I don't know if it was going into parasitic oscillation or the base-emitter junction had somehow become damaged in a way that caused this behavior. Replacing the transistor fixed it. Very strange behavior for a simple semiconductor.
 

FRITZHID

Flashlight Enthusiast
Joined
Jun 20, 2011
Messages
2,500
Location
Icelandic wastelands of Monico, WI
In my work life, I do some validation testing of electronics. Some of this includes slowly ramping up the supply voltage. There have been cases where semiconductors latch up due to junctions not getting reverse biased at the right time. Think of the substrate in an IC that is doped material, but must be reverse biased in order to avoid conducting. Well, once it starts conducting, it's not going to stop until you remove power.

I've also seen this sort of thing during ESD testing, for what it's worth. It's amazing what sort of weird stuff happens when you poke and taunt the electronics in just the right way. :)

++2 agreed. They do some very odd things at times. I've had some electronics that work fine between 120v-125v but if dropped below 119v all sorts of things would happen, @ only 1v difference. It can be terribly hard to predict how they'll behave under certain conditions.
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
++2 agreed. They do some very odd things at times. I've had some electronics that work fine between 120v-125v but if dropped below 119v all sorts of things would happen, @ only 1v difference. It can be terribly hard to predict how they'll behave under certain conditions.

It's possible Philips put in a latch circuit to shut them down to prevent overheating from operating too long below a certain voltage. I guarantee they would have gone through brown-out testing and given that two did the same thing, it seems more like a planned operation than anything else.

Semiman
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
It's possible Philips put in a latch circuit to shut them down to prevent overheating from operating too long below a certain voltage. I guarantee they would have gone through brown-out testing and given that two did the same thing, it seems more like a planned operation than anything else.

Semiman

Would there be an advantage to using a latching circuit instead of a low voltage detector (with a little hysteresis)?

I agree that validation testing should have caught something like this, but I also recall an IC supplier updating one of their products, with the result that their IC was prone to latch-up. Since this was just an IC update, we generally don't do a full series of validation tests. I don't recall if we ran the power-ramp-up/down test on it or not... it was pretty hard to duplicate, but it was there. The IC supplier wasn't aware of the problem, and had to be shown that it existed.

Most IC suppliers are good about telling their customers when they change the IC (usually a die shrink). I wonder how many of their customers run full electrical validation tests with the updated part?
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
Latching shotdown could be timed. I have done that. Allow operation at a low voltage for a period of time then shutdown.
 

more_vampires

Flashlight Enthusiast
Joined
Nov 20, 2014
Messages
3,475
I've had PC PSUs get funky when rapidly power cycled in short order. The weirdest one became unstable, then "calmed down" when allowed to rest for a few days, never a problem with it after that and it even ran stable. I then later duplicated the problem! It then "calmed down" again. A duplicatable "temporary bricking."

It was one of the weirdest glitches I've seen, an old head told me not to rapidly cycle a PSU and let it set for 1 minute after a full shutdown before attempting a cold start. Guess he was right?

Betting if I hard discharged all the caps in that PSU that it would have come back to life right away.
 

PhotonWrangler

Flashaholic
Joined
Oct 19, 2003
Messages
14,432
Location
In a handbasket
Yeah it does seem like it might be one of those undocumented "features" since both lamps acted exactly the same. I haven't had a chance to dig out the variac yet but when I do I'll post the results.
 

PhotonWrangler

Flashaholic
Joined
Oct 19, 2003
Messages
14,432
Location
In a handbasket
Recently one of these alien-head bulbs started blinking again. I took the lenses and top off and re-seated the connectors, which has corrected the blinking for now. However -

1) The plastic connector bodies have become extremely brittle, and I chipped one of them in a way that it broke off some insulation between the red & black wires. Added a dab of silicone for insulation.

2) I'm noticing a 60hz flicker in the light output of this bulb, so the electrolytic caps must be failing.

It's been running for over a year, maybe 6 hours per day. I'm going to start writing in-service dates on LED bulbs before I install them.
 
Top